What AI 2027 Means for Your Job, Creativity, and the Future of Work
AI is already good and it’s about to become more powerful in the next two-five years. So the question is - how do we stay valuable when AI is sprinting 24/7?
I'm launching this blog as a response to being overwhelmed by the rapid pace of AI development. As Deb Liu said in one of her recent posts, we can't fight it anymore - so in the long run, the winners will be those who learn to harness AI. So be it.
In this blog, I’ll be exploring various AI tools - how to use them in product management and in life to maximize my value per hour. I’ll share prompts and workflows of several tools connected to each other, and reflect on what I’ve learned and what I’m reading on the topic. I’ll also spin up a “productivity dashboard” to track time optimizations!
But before we dive in, let’s take a look at what kind of future we all are headed toward.
If you haven't yet read the AI-2027 report (Report), I highly recommend spending an hour to look through. It’s a gripping, sometimes unsettling, look into where this all might be going. Below are my key takeaways but TL;DR is this:
AI is already good and it’s about to become more powerful in the next two-five years. Routine tasks like taking meeting notes or writing good enough documents will be fully automated. Even today, there’s not much sense to learn how to code or write SQL - AI does that job for you. What we’re left with is creativity and human judgement.
Key takeaways
1. AI agents will likely morph into something else
Have you ever used a chat bot? It can answer pre-programmed questions, and perform some basic operations. A couple of years ago, ChatGPT could help you turn bullet points into emails and vice versa, fix grammar, that sort of thing. But agents are different. They can perform much more complex tasks like generating a working piece of code based on your text instructions.The Report, however, says that agents have seen limited success so far, mostly because they're not reliable and cosy. My take is that the concept is also not easy to understand and integrate into your day-to-day life if you're not the one who actually developed an agent.
Action item: Spend some time playing with agents to understand what they can (and can’t) do. The shift is coming.
2. Risk of fraud and malicious use will grow
The Report says that while big tech giants will continue advocating for AI agents usefulness and harmless, i.e. they refuse to take part in any dangerous activity, the nature of AI (they're massive neural networks) suggests that they simply can't be programmed with specific “do” and “don't” hard coded into their behavior. Instead, they optimize for rewards from a human being interacting with them which means they can deceive, cut corners and manipulate. And while researchers would likely try to identify cases where models seem to deviate from the spec, fraud, attacks and malicious use of models would still likely become more sophisticated thanks to AI capabilities.
3. AI will take some jobs
AI will keep improving at systematic, rule-based tasks - coding, analysis, short-form writing. It’s already powerful at writing SQL queries, building websites, summarizing meetings, and drafting strategy docs. Two to five years from now, those workflows will be fully automated and won’t require a human.
Today, many companies out there are saying that AI won’t replace your job, it will create even more jobs, just a different nature: e.g. Product Engineer, AI Reviewer, etc.. I actually think this is not true. The Report agrees:
“The job market for junior software engineers is in turmoil: the AIs can do everything taught by a CS degree, but people who know how to manage and quality-control teams of AIs are making a killing.”
The truth is, AI never sleeps, doesn’t burn out, and doesn’t need a vacation.
4. Data will be the new oil
Let’s be real - data is already everything. If your company has access to data from thousands of customers, you can generate an enormous number of insights to improve your business and differentiate from competitors. With AI, shifts towards data importance will become even more prominent. The deal is that AI needs massive amounts of training data, especially for more complex, long-horizon tasks (like managing a coffee shop or supply chain). According to the Report, some companies will start paying humans to generate custom training data. Sounds wild, I know!
5. The most advanced AI will be private
It is likely the case today, but it’s about to accelerate. The most advanced AI-models and capabilities will be classified, and siloed inside governments, military programs, and private labs with serious security protocols.
So the question is - how do we stay valuable when AI is sprinting 24/7?